Spatial and temporal scales of distribution and demography in breeding songbirds: implications of habitat fragmentation and restoration

Abstract

<p>Tallgrass prairie has declined throughout the midwestern United States during the past two centuries, and migratory birds breeding in these habitats have also experienced precipitous population declines. One conservation strategy used to mitigate the effects of habitat loss and fragmentation is habitat restoration. I studied how both habitat fragmentation and restoration affect songbird populations breeding in grassland and wetland habitats in northern Iowa, 1999--2002. Most grassland birds tended to be less abundant near edges, yet birds avoided woodland edges more so than other types of edges. Edge avoidance could not be explained by changes in habitat structure. For Bobolinks (Dolichonyx oryzivorus), edge avoidance also increased near corners of sites, where multiple edges converged. These local patterns of edge avoidance were consistent with observed patterns of area sensitivity when scaling patterns up to fragmented landscapes using simulation modeling in neutral (randomized) landscapes. Moreover, regression modeling at landscape scales suggested that edge density metrics were better than landscape compositional metrics at explaining bird density within patches. Habitat restoration provided breeding habitat for many bird species, in which birds tended to occur at high densities in restored patches relative to other land cover types in the landscape. However, demographic data revealed a more complex pattern. In restored wetlands, reproduction was tied closely to interannual climate variation. In particular, nest predation was negatively correlated with water depth in wetlands. In restored grasslands, nest success tended to be relatively low for most species. Population projection models suggested that for Dickcissels (Spiza americana) and, to a lesser extent, Bobolinks, population growth rates were not high enough to be sustainable without immigration into the area (lambda < 1). In addition, population growth was most sensitive to adult survival. Sensitivity to nest predation was moderate and dependent on estimates of survival, while growth was less sensitive to brood parasitism and juvenile survival. Results from this study have improved our understanding of the role of edge effects in generating spatial distributions in landscapes, how restoration ultimately affects avian populations in the Midwest, and it provides a framework for understanding songbird dynamics in fragmented landscapes.</p>